Neighbourhood Watch Victoria survey analysis¶
In [1]:
#!pip install geopandas
In [2]:
import pandas as pd
import geopandas as gpd
import matplotlib.pyplot as plt
import seaborn as sns
sns.set(style="darkgrid")
%matplotlib inline
In [3]:
#Read the data
df = pd.read_csv("Copy of HSIMP VIC Quiz Results 030620_2.csv",sep =',')
df.head()
Out[3]:
Data Preprocessing¶
In [4]:
#Drop unneccesary columns
df = df.drop(['Name', 'Date'], axis = 1)
In [5]:
#Check unique values
for col in df:
print('Unique values of ' + col)
print(df[col].unique())
In [6]:
#Replace NA
for col in df:
print('\nAfter transformation')
df[col] = df[col].fillna('No Response')
print(df[col].unique())
print('')
In [7]:
print('\nAfter transformation')
df['Postcode'] = df['Postcode'].replace({'Rooney': 0, 'mer': 0,'No Response': 0, 'Thorne':0, 'Grivas':0,'3084he': 3084})
print(df['Postcode'].unique())
In [8]:
#Convert to numeric
df['Postcode'] = pd.to_numeric(df['Postcode'],errors='coerce')
In [9]:
#load suburb, lat long data
df1 = pd.read_csv("location.csv",sep =',')
In [10]:
df1.head()
Out[10]:
In [11]:
#Merge using postcode as key
u = df.assign(cnt=df.groupby('Postcode').cumcount())
v = df1.assign(cnt=df1.groupby('Postcode').cumcount())
In [12]:
v = v.drop_duplicates(subset='Postcode', keep='first')
In [13]:
print(v.shape)
v.head()
Out[13]:
In [14]:
df_new = u.merge(v, on=['Postcode'], how='left')
print(df_new.shape)
df_new.head()
Out[14]:
In [15]:
df_new.isnull().sum()
Out[15]:
In [16]:
#!pip install descartes
Load shape file of Victoria¶
In [17]:
nb = 'VIC_LOCALITY_POLYGON_shp.shp'
regions = gpd.read_file(nb)
regions.head()
Out[17]:
In [18]:
regions.rename(columns={'VIC_LOCA_2': 'suburb'}, inplace=True)
In [19]:
regions.head()
Out[19]:
In [20]:
regions.shape
Out[20]:
In [21]:
#Load Victoria map
regions.plot(figsize=(40,20))
plt.show()
In [22]:
suburb = df_new.loc[:,'suburb']
rating = df_new.loc[:,'How safe do you feel when you are in your home?']
suburb = pd.DataFrame({'suburb': suburb,'rating':rating})
suburb['suburb'] = suburb['suburb'].str.lower()
regions['suburb'] =regions['suburb'].str.lower()
In [23]:
# Join the dataset with the available mapping files.
# We will fill in the neighbourhoods which were not present in the dataset with 0.
# this way we will still see the neighbourhood in the map, but no housing will be shown on it.
merged = regions.set_index('suburb').join(suburb.set_index('suburb'))
merged = merged.reset_index()
merged = merged.fillna(0)
merged[['suburb', 'geometry', 'rating']].head()
Out[23]:
In [24]:
merged['rating'] = pd.to_numeric(merged['rating'],errors='coerce')
Overall safety ratings across Victoria¶
In [25]:
fig, ax = plt.subplots(1, figsize=(40, 20))
ax.axis('off')
ax.set_title('Heat Map of Safety Ratings in Victoria, Australia', fontdict={'fontsize': '40', 'fontweight' : '3'})
color = 'YlGn'
vmin, vmax = 0, 10
sm = plt.cm.ScalarMappable(cmap=color, norm=plt.Normalize(vmin=vmin, vmax=vmax))
sm._A = []
cbar = fig.colorbar(sm)
cbar.ax.tick_params(labelsize=20)
# actually plot the map
# we will only annotate the plot for neighbourhoods with more than 50 houses sold
merged.plot('rating', cmap=color, linewidth=0.8, ax=ax, edgecolor='0.8', figsize=(40,20))
for idx, row in merged.iterrows():
if(row['rating'] > 50):
plt.annotate(text=row['suburb'], xy=(row['geometry'], row['geometry']),
horizontalalignment='center', fontsize='large', color='black', wrap=True)
plt.show()
fig.savefig('plot1.pdf')
Safest suburbs with safety rating of 10¶
In [26]:
merged_copy = merged.copy()
In [27]:
print('\nAfter transformation')
merged_copy['rating'] = merged_copy['rating'].replace({1.0: 0, 2.0:0, 3.0:0,4.0:0,5.0:0,6.0:0, 7.0:0, 8.0:0,9.0:0})
print(merged_copy['rating'].unique())
print('')
In [28]:
merged.shape
Out[28]:
In [29]:
fig, ax = plt.subplots(1, figsize=(40, 20))
ax.axis('off')
ax.set_title('Heat Map of Highest Safety Rating of 10 in Victoria, Australia', fontdict={'fontsize': '40', 'fontweight' : '3'})
color = 'YlGn'
vmin, vmax = 0, 10
sm = plt.cm.ScalarMappable(cmap=color, norm=plt.Normalize(vmin=vmin, vmax=vmax))
sm._A = []
cbar = fig.colorbar(sm)
cbar.ax.tick_params(labelsize=20)
# actually plot the map
# we will only annotate the plot for neighbourhoods with more than 50 houses sold
merged.plot('rating', cmap=color, linewidth=0.8, ax=ax, edgecolor='0.8', figsize=(40,20))
for idx, row in merged.iterrows():
if(row['rating'] > 50):
plt.annotate(text=row['suburb'], xy=(row['geometry'], row['geometry']),
horizontalalignment='center', fontsize='large', color='black', wrap=True)
plt.show()
fig.savefig('plot2.pdf')
Moderately safe suburbs with safety rating between 9 to 7¶
In [30]:
merged_copy2 = merged.copy()
In [31]:
print('\nAfter transformation')
merged_copy2['rating'] = merged_copy2['rating'].replace({1.0: 0, 2.0:0, 3.0:0,4.0:0,5.0:0,6.0:0, 10.0:0})
print(merged_copy2['rating'].unique())
print('')
In [32]:
fig, ax = plt.subplots(1, figsize=(40, 20))
ax.axis('off')
ax.set_title('Heat Map of Safety Ratings between 9 to 7 in Victoria, Australia', fontdict={'fontsize': '40', 'fontweight' : '3'})
color = 'YlGn'
vmin, vmax = 0, 10
sm = plt.cm.ScalarMappable(cmap=color, norm=plt.Normalize(vmin=vmin, vmax=vmax))
sm._A = []
cbar = fig.colorbar(sm)
cbar.ax.tick_params(labelsize=20)
# actually plot the map
# we will only annotate the plot for neighbourhoods with more than 50 houses sold
merged.plot('rating', cmap=color, linewidth=0.8, ax=ax, edgecolor='0.8', figsize=(40,20))
for idx, row in merged.iterrows():
if(row['rating'] > 50):
plt.annotate(text=row['suburb'], xy=(row['geometry'], row['geometry']),
horizontalalignment='center', fontsize='large', color='black', wrap=True)
plt.show()
fig.savefig('plot2.pdf')
Unsafe/least safe suburbs with safety rating between 6 to 1¶
In [33]:
merged_copy3 = merged.copy()
In [34]:
print('\nAfter transformation')
merged_copy3['rating'] = merged_copy3['rating'].replace({1.0: 0, 2.0:0, 3.0:0,4.0:0,5.0:0,6.0:0, 10.0:0})
print(merged_copy3['rating'].unique())
print('')
In [35]:
fig, ax = plt.subplots(1, figsize=(40, 20))
ax.axis('off')
ax.set_title('Heat Map of Lowest Safety Ratings between 6 to 1 in Victoria, Australia', fontdict={'fontsize': '40', 'fontweight' : '3'})
color = 'Oranges'
vmin, vmax = 0, 6
sm = plt.cm.ScalarMappable(cmap=color, norm=plt.Normalize(vmin=vmin, vmax=vmax))
sm._A = []
cbar = fig.colorbar(sm)
cbar.ax.tick_params(labelsize=20)
# actually plot the map
# we will only annotate the plot for neighbourhoods with more than 50 houses sold
merged.plot('rating', cmap=color, linewidth=0.8, ax=ax, edgecolor='0.8', figsize=(40,20))
for idx, row in merged.iterrows():
if(row['rating'] > 50):
plt.annotate(text=row['suburb'], xy=(row['geometry'], row['geometry']),
horizontalalignment='center', fontsize='large', color='black', wrap=True)
plt.show()
Overall safety ratings¶
In [36]:
suburb = df_new.loc[:,'suburb']
suburb.head()
Out[36]:
In [37]:
rating = df_new.loc[:,'How safe do you feel when you are in your home?']
rating.head()
Out[37]:
In [38]:
postcode = df_new.loc[:,'Postcode']
postcode.head()
Out[38]:
In [39]:
suburb = pd.DataFrame({'suburb': suburb,'rating':rating, 'postcode': postcode})
suburb.head()
Out[39]:
In [40]:
suburb['suburb'] = suburb['suburb'].str.lower()
In [41]:
regions['suburb'] =regions['suburb'].str.lower()
In [42]:
print('\nAfter transformation')
suburb['rating'] = suburb['rating'].replace({'No Response': 0})
print(suburb['rating'].unique())
print('')
In [43]:
print(suburb.shape)
print(regions.shape)
In [44]:
a = suburb.assign(cnt=suburb.groupby('suburb').cumcount())
b = regions.assign(cnt=regions.groupby('suburb').cumcount())
In [45]:
merged1 =a.merge(b, on=['suburb','cnt'], how='left')
print(merged1.shape)
merged1.head(2)
Out[45]:
In [46]:
#Drop no postcode or wrong postcodes
merged1.dropna(subset=['suburb'], inplace=True)
print(merged1.shape)
In [47]:
merged1['rating'].value_counts().sort_index()
Out[47]:
In [48]:
df_piechart = merged1['rating'].value_counts().sort_index()
In [49]:
import matplotlib.pyplot as plt
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = df_piechart.index
sizes = df_piechart.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.0, 1),
bbox_transform=fig1.transFigure
)
plt.title("Percentage of Safety Ratings (in %)")
Out[49]:
In [50]:
import numpy as np
fig, ax1 = plt.subplots()
#Plot 1
#Treatment feature
df_treatment = merged1['rating'].value_counts().sort_index()
objects = df_treatment.index
y_pos = np.arange(len(objects))
performance = df_treatment.values
ax1.bar(y_pos, performance, align='center', alpha=0.7, color='G')
ax1.set_xticks(y_pos)
ax1.set_xticklabels(df_treatment.index)
ax1.set_ylabel('Count')
ax1.set_xlabel('Safety Ratings')
ax1.set_title('Total count of Safety Ratings')
ax1.set_facecolor('white')
ax1.grid(color='#95a5a6', linewidth=.8, axis='y', alpha=0.3)
for p in ax1.patches:
height = p.get_height()
ax1.text(p.get_x()+p.get_width()/2., height + 0.1,height ,ha="center", fontsize=9)
plt.show()
Safest suburbs with safety rating of 10¶
In [51]:
merged1_copy = merged1.copy()
In [52]:
len(merged1_copy['suburb'].unique())
Out[52]:
In [53]:
merged1_copy.shape
Out[53]:
In [54]:
#Filter 10 safety rating
safe_subs = merged1_copy[(merged1_copy['rating'] == 10)]
print(safe_subs.shape)
safe_subs.head()
Out[54]:
In [55]:
len(safe_subs['suburb'].unique())
Out[55]:
In [56]:
safe_subs[['suburb', 'rating', 'postcode']]
Out[56]:
In [57]:
slide2 = safe_subs[['suburb', 'rating', 'postcode']].sort_values(by = 'suburb', ascending = True)
In [58]:
slide2 = slide2.drop_duplicates(subset='suburb', keep='first')
slide2_2 =slide2.iloc[34:68]
slide2_3 =slide2.iloc[68:127]
In [59]:
import six
def render_mpl_table(data, col_width=3.0, row_height=0.625, font_size=14,
header_color='lightseagreen', row_colors=['#f1f1f2', 'w'], edge_color='w',
bbox=[0, 0, 1, 1], header_columns=0,
ax=None, **kwargs):
if ax is None:
size = (np.array(data.shape[::-1]) + np.array([0, 1])) * np.array([col_width, row_height])
fig, ax = plt.subplots(figsize=size)
ax.axis('off')
mpl_table = ax.table(cellText=data.values, bbox=bbox, colLabels=data.columns, **kwargs)
mpl_table.auto_set_font_size(False)
mpl_table.set_fontsize(font_size)
for k, cell in six.iteritems(mpl_table._cells):
cell.set_edgecolor(edge_color)
if k[0] == 0 or k[1] < header_columns:
cell.set_text_props(weight='bold', color='w')
cell.set_facecolor(header_color)
else:
cell.set_facecolor(row_colors[k[0]%len(row_colors) ])
return ax
render_mpl_table(slide2_3, header_columns=0, col_width=2.5)
Out[59]:
Type of Home¶
In [60]:
slide3 = df_new.rename(columns={'How safe do you feel when you are in your home?': 'rating'})
In [61]:
print('\nAfter transformation')
slide3['rating'] = slide3 ['rating'].replace({'No Response': 0})
print(slide3['rating'].unique())
print('')
In [62]:
#Drop no postcode or wrong postcodes
slide3.dropna(subset=['suburb'], inplace=True)
print(slide3.shape)
In [63]:
#Filter 10 again
safe_subs3 = slide3[(slide3['rating'] == 10)]
print(safe_subs3.shape)
safe_subs3.head()
Out[63]:
In [64]:
safe_subs3['What type of home do you live in?'].value_counts().sort_values()
Out[64]:
In [65]:
df_piechart2 = safe_subs3['What type of home do you live in?'].value_counts().sort_values()
In [66]:
df_piechart2 = safe_subs3['What type of home do you live in?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = df_piechart2.index
sizes = df_piechart2.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of Home (in %)")
Out[66]:
Type of Doors¶
In [67]:
df_piechart3 = safe_subs3['Which of the following external doors does your home have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = df_piechart3.index
sizes = df_piechart3.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of External Doors (in %)")
Out[67]:
Type of Fences¶
In [68]:
df_piechart4 = safe_subs3['What fences does your property have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = df_piechart4.index
sizes = df_piechart4.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.9, 1),
bbox_transform=fig1.transFigure
)
plt.title("Fences in the House(in %)")
Out[68]:
Security alarm at home¶
In [69]:
df_piechart4 = safe_subs3['Does your home have any security systems?'].value_counts().sort_index()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = df_piechart4.index
sizes = df_piechart4.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
theme = plt.get_cmap('RdYlGn')
#theme = plt.get_cmap('hot')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Does your home have any\n security systems? (in %)")
Out[69]:
House security measures¶
In [70]:
doorslocked = safe_subs3['I keep all my doors and windows locked when I’m at home or in my yard'].value_counts().sort_index()
doorslockedperc = np.true_divide(doorslocked.values, doorslocked.sum())*100
In [71]:
skylightlocked = safe_subs3['Any manholes or skylights are kept locked and/or bolted from the inside'].value_counts().sort_index()
In [72]:
computers =safe_subs3['Computers, televisions and other electronics not visible from outside'].value_counts().sort_index()
computersperc =np.true_divide(computers.values, computers.sum())*100
In [73]:
doorseen = safe_subs3['Front door and entry can be seen from street'].value_counts().sort_index()
doorseenperc =np.true_divide(doorseen.values, doorseen.sum())*100
In [74]:
housenum = safe_subs3['House number is clearly displayed'].value_counts().sort_index()
housenumperc =np.true_divide(housenum.values, housenum.sum())*100
In [75]:
housenumvis = safe_subs3['House number is visible at night'].value_counts().sort_index()
housenumvisperc =np.true_divide(housenumvis.values, housenumvis.sum())*100
In [76]:
gard = safe_subs3['Garden kept short and neat'].value_counts().sort_index()
gardperc =np.true_divide(gard.values, gard.sum())*100
In [77]:
camera= safe_subs3['Security camera faces entry'].value_counts().sort_index()
camperc = np.true_divide(camera.values, camera.sum())*100
In [78]:
dog= safe_subs3['I have a dog'].value_counts().sort_index()
dogperc = np.true_divide(dog.values, dog.sum())*100
In [79]:
value= safe_subs3['Valuables kept in a secured, locked safe located out of sight.'].value_counts().sort_index()
valueperc = np.true_divide(value.values, value.sum())*100
In [80]:
sensorlight= safe_subs3['I have sensor lighting or timed lighting in my house for when I’m not at home'].value_counts().sort_index()
sensorlightperc = np.true_divide(sensorlight.values, sensorlight.sum())*100
In [81]:
col_names = ["No",'Yes']
results = { 'Front door seen from street': doorseenperc,
'House number is clearly displayed': housenumperc,
'House number is visible at night': housenumvisperc,
"Doors and windows locked": doorslockedperc,
'Electronics out of sight': computersperc,
'Garden kept short and neat': gardperc,
'Sensor/timed lighting when not at home': sensorlightperc,
'Security camera at home': camperc,
'Dog at home': dogperc,
'Valuables kept in a secured & locked safe': valueperc}
def survey(results, col_names):
labels = list(results.keys())
data = np.array(list(results.values()))
data_cum = data.cumsum(axis=1)
category_colors = plt.get_cmap('RdYlGn')(np.linspace(0.1, 0.90, data.shape[1]))
fig, ax = plt.subplots(figsize=(9.9, 3.7))
ax.invert_yaxis()
ax.xaxis.set_visible(False)
ax.set_xlim(0, np.sum(data, axis=1).max())
for i, (colname, color) in enumerate(zip(col_names, category_colors)):
widths = data[:, i]
starts = data_cum[:, i] - widths
ax.barh(labels, widths, left=starts, height=0.9, label=colname, color=color)
xcenters = starts + widths / 2
r, g, b, _ = color
text_color = 'black'
for y, (x, c) in enumerate(zip(xcenters, widths)):
ax.text(x, y, str(int(c)), ha='center', va='center',
color=text_color)
ax.legend(ncol=len(col_names), bbox_to_anchor=(0.42, -0.15),
loc='lower left', fontsize='small')
return fig, ax
survey(results, col_names)
plt.title("House security measures (in %)")
fig.tight_layout()
plt.show()
fig.savefig('plot5.pdf')
Moderately safe suburbs with safety rating between 9 to 7¶
In [82]:
#Filter 9 to 7 safety rating
msafe_subs = merged1_copy[merged1_copy.rating.isin([7,8,9])]
print(msafe_subs.shape)
In [83]:
len(msafe_subs['suburb'].unique())
Out[83]:
In [84]:
table3 =msafe_subs[['suburb', 'rating', 'postcode']].sort_values(by = 'suburb', ascending = True)
In [85]:
table3 = table3.drop_duplicates(subset='suburb', keep='first')
table3_1 =table3.iloc[0:34]
table3_2 =table3.iloc[34:68]
table3_3 =table3.iloc[68:120]
In [86]:
import six
def render_mpl_table(data, col_width=3.0, row_height=0.625, font_size=14,
header_color='lightseagreen', row_colors=['#f1f1f2', 'w'], edge_color='w',
bbox=[0, 0, 1, 1], header_columns=0,
ax=None, **kwargs):
if ax is None:
size = (np.array(data.shape[::-1]) + np.array([0, 1])) * np.array([col_width, row_height])
fig, ax = plt.subplots(figsize=size)
ax.axis('off')
mpl_table = ax.table(cellText=data.values, bbox=bbox, colLabels=data.columns, **kwargs)
mpl_table.auto_set_font_size(False)
mpl_table.set_fontsize(font_size)
for k, cell in six.iteritems(mpl_table._cells):
cell.set_edgecolor(edge_color)
if k[0] == 0 or k[1] < header_columns:
cell.set_text_props(weight='bold', color='w')
cell.set_facecolor(header_color)
else:
cell.set_facecolor(row_colors[k[0]%len(row_colors) ])
return ax
render_mpl_table(table3_3, header_columns=0, col_width=2.5)
Out[86]:
In [87]:
#Filter 7 to 9 safety rating
midsafe_subs =slide3[slide3.rating.isin([7,8,9])]
print(midsafe_subs.shape)
midsafe_subs.head()
Out[87]:
In [88]:
len(midsafe_subs['suburb'].unique())
Out[88]:
Type of Home¶
In [89]:
midsafe_subs['What type of home do you live in?'].value_counts().sort_values()
Out[89]:
In [90]:
mdf_piechart1 = midsafe_subs['What type of home do you live in?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = mdf_piechart1.index
sizes = mdf_piechart1.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of Home (in %)")
Out[90]:
Type of Doors¶
In [91]:
mdf_piechart2 = midsafe_subs['Which of the following external doors does your home have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = mdf_piechart2.index
sizes = mdf_piechart2.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.9, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of External Doors (in %)")
Out[91]:
Type of Fence¶
In [92]:
mdf_piechart3 = midsafe_subs['What fences does your property have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = mdf_piechart3.index
sizes = mdf_piechart3.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.9, 1),
bbox_transform=fig1.transFigure
)
plt.title("Fences in the House(in %)")
Out[92]:
Security alarm at home¶
In [93]:
mdf_piechart4 = midsafe_subs['Security camera faces entry'].value_counts().sort_index()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = mdf_piechart4.index
sizes = mdf_piechart4.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Does your home have any\n security systems? (in %)")
Out[93]:
House security measures¶
In [94]:
doorslockedm = midsafe_subs['I keep all my doors and windows locked when I’m at home or in my yard'].value_counts().sort_index()
doorslockedpercm = np.true_divide(doorslockedm.values, doorslockedm.sum())*100
doorslockedpercm
Out[94]:
In [95]:
computersm =midsafe_subs['Computers, televisions and other electronics not visible from outside'].value_counts().sort_index()
computerspercm =np.true_divide(computersm.values, computersm.sum())*100
computerspercm
Out[95]:
In [96]:
doorseenm = midsafe_subs['Front door and entry can be seen from street'].value_counts().sort_index()
doorseenpercm =np.true_divide(doorseenm.values, doorseenm.sum())*100
doorseenpercm
Out[96]:
In [97]:
housenumm = midsafe_subs['House number is clearly displayed'].value_counts().sort_index()
housenumpercm =np.true_divide(housenumm.values, housenumm.sum())*100
housenumpercm
Out[97]:
In [98]:
housenumvism = midsafe_subs['House number is visible at night'].value_counts().sort_index()
housenumvispercm =np.true_divide(housenumvism.values, housenumvism.sum())*100
housenumvispercm
Out[98]:
In [99]:
gardm = midsafe_subs['Garden kept short and neat'].value_counts().sort_index()
gardpercm =np.true_divide(gardm.values, gardm.sum())*100
gardpercm
Out[99]:
In [100]:
cameram= midsafe_subs['Security camera faces entry'].value_counts().sort_index()
campercm = np.true_divide(cameram.values, cameram.sum())*100
campercm
Out[100]:
In [101]:
dogm= midsafe_subs['I have a dog'].value_counts().sort_index()
dogpercm = np.true_divide(dogm.values, dogm.sum())*100
dogpercm
Out[101]:
In [102]:
sensorlightm= midsafe_subs['I have sensor lighting or timed lighting in my house for when I’m not at home'].value_counts().sort_index()
sensorlightpercm = np.true_divide(sensorlightm.values, sensorlightm.sum())*100
sensorlightpercm
Out[102]:
In [103]:
valuem= midsafe_subs['Valuables kept in a secured, locked safe located out of sight.'].value_counts().sort_index()
valuepercm = np.true_divide(valuem.values, valuem.sum())*100
valuepercm
Out[103]:
In [104]:
col_names = ["No",'Yes']
results = { 'Front door seen from street': doorseenpercm,
'House number is clearly displayed': housenumpercm,
'House number is visible at night': housenumvispercm,
"Doors and windows locked": doorslockedpercm,
'Electronics out of sight': computerspercm,
'Garden kept short and neat': gardpercm,
'Sensor/timed lighting when not at home': sensorlightpercm,
'Security camera at home': campercm,
'Dog at home': dogpercm,
'Valuables kept in a secured & locked safe': valuepercm}
def survey(results, col_names):
labels = list(results.keys())
data = np.array(list(results.values()))
data_cum = data.cumsum(axis=1)
category_colors = plt.get_cmap('RdYlGn')(np.linspace(0.1, 0.90, data.shape[1]))
fig, ax = plt.subplots(figsize=(9.9, 3.7))
ax.invert_yaxis()
ax.xaxis.set_visible(False)
ax.set_xlim(0, np.sum(data, axis=1).max())
for i, (colname, color) in enumerate(zip(col_names, category_colors)):
widths = data[:, i]
starts = data_cum[:, i] - widths
ax.barh(labels, widths, left=starts, height=0.9, label=colname, color=color)
xcenters = starts + widths / 2
r, g, b, _ = color
text_color = 'black'
for y, (x, c) in enumerate(zip(xcenters, widths)):
ax.text(x, y, str(int(c)), ha='center', va='center',
color=text_color)
ax.legend(ncol=len(col_names), bbox_to_anchor=(0.42, -0.15),
loc='lower left', fontsize='small')
return fig, ax
survey(results, col_names)
plt.title("House security measures (in %)")
fig.tight_layout()
plt.show()
Unsafe/least safe suburbs with safety rating between 6 to 1¶
In [105]:
#Filter 6 to 1 safety rating
nsafe_subs = merged1_copy[merged1_copy.rating.isin([1,2,3,4,5,6])]
print(nsafe_subs.shape)
nsafe_subs.head()
Out[105]:
In [106]:
len(nsafe_subs['suburb'].unique())
Out[106]:
In [107]:
table2 = nsafe_subs[['suburb', 'rating', 'postcode']].sort_values(by = 'suburb', ascending = True)
In [108]:
table2 = table2.drop_duplicates(subset='suburb', keep='first')
table2_1 =table2.iloc[0:34]
table2_2 =table2.iloc[34:68]
table2_3 =table2.iloc[68:120]
In [109]:
import six
def render_mpl_table(data, col_width=3.0, row_height=0.625, font_size=14,
header_color='lightseagreen', row_colors=['#f1f1f2', 'w'], edge_color='w',
bbox=[0, 0, 1, 1], header_columns=0,
ax=None, **kwargs):
if ax is None:
size = (np.array(data.shape[::-1]) + np.array([0, 1])) * np.array([col_width, row_height])
fig, ax = plt.subplots(figsize=size)
ax.axis('off')
mpl_table = ax.table(cellText=data.values, bbox=bbox, colLabels=data.columns, **kwargs)
mpl_table.auto_set_font_size(False)
mpl_table.set_fontsize(font_size)
for k, cell in six.iteritems(mpl_table._cells):
cell.set_edgecolor(edge_color)
if k[0] == 0 or k[1] < header_columns:
cell.set_text_props(weight='bold', color='w')
cell.set_facecolor(header_color)
else:
cell.set_facecolor(row_colors[k[0]%len(row_colors) ])
return ax
render_mpl_table(table2_3, header_columns=0, col_width=2.5)
Out[109]:
In [110]:
#Filter 1 to 6 safety rating
notsafe_subs =slide3[slide3.rating.isin([1,2,3,4,5,6])]
print(notsafe_subs.shape)
notsafe_subs.head()
Out[110]:
In [111]:
table2 = notsafe_subs.drop_duplicates(subset='suburb', keep='first')
In [112]:
len(notsafe_subs['suburb'].unique())
Out[112]:
Type of Home¶
In [113]:
notsafe_subs['What type of home do you live in?'].value_counts().sort_values()
Out[113]:
In [114]:
ndf_piechart1 = notsafe_subs['What type of home do you live in?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = ndf_piechart1.index
sizes = ndf_piechart1.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of Home (in %)")
Out[114]:
Type of Doors¶
In [115]:
ndf_piechart2 = notsafe_subs['Which of the following external doors does your home have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = ndf_piechart2.index
sizes = ndf_piechart2.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Type of External Doors (in %)")
Out[115]:
Type of Fences¶
In [116]:
ndf_piechart3 = notsafe_subs['What fences does your property have?'].value_counts().sort_values()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = ndf_piechart3.index
sizes = ndf_piechart3.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.9, 1),
bbox_transform=fig1.transFigure
)
plt.title("Fences in the House(in %)")
Out[116]:
Security alarms at home¶
In [117]:
ndf_piechart5 = notsafe_subs['Security camera faces entry'].value_counts().sort_index()
def autopct_generator(limit):
"""Remove percent on small slices."""
def inner_autopct(pct):
return ('%.2f%%' % pct) if pct > limit else ''
return inner_autopct
labels = ndf_piechart5.index
sizes = ndf_piechart5.values
NUM_COLORS = len(sizes)
fig1, ax1 = plt.subplots(figsize=(6, 5))
#theme = plt.get_cmap('YlGn')
theme = plt.get_cmap('RdYlGn')
ax1.set_prop_cycle("color", [theme(1. * i / len(sizes)) for i in range(len(sizes))])
box = ax1.get_position()
ax1.set_position([box.x0, box.y0, box.width * 1.3, box.height])
_, _, autotexts = ax1.pie(
sizes, autopct=autopct_generator(7), startangle=90, radius=1.8 * 1000)
for autotext in autotexts:
autotext.set_weight('bold')
ax1.axis('equal')
total = sum(sizes)
plt.legend(
loc='upper left',
labels=['%s, %1.1f%%' % (
l, (float(s) / total) * 100) for l, s in zip(labels, sizes)],
prop={'size': 11},
bbox_to_anchor=(0.85, 1),
bbox_transform=fig1.transFigure
)
plt.title("Does your home have any\n security systems? (in %)")
Out[117]:
House security measures¶
In [118]:
doorslockedn = notsafe_subs['I keep all my doors and windows locked when I’m at home or in my yard'].value_counts().sort_index()
doorslockedpercn = np.true_divide(doorslockedn.values, doorslockedn.sum())*100
doorslockedpercn
Out[118]:
In [119]:
computersn =notsafe_subs['Computers, televisions and other electronics not visible from outside'].value_counts().sort_index()
computerspercn =np.true_divide(computersn.values, computersn.sum())*100
computerspercn
Out[119]:
In [120]:
doorseenn = notsafe_subs['Front door and entry can be seen from street'].value_counts().sort_index()
doorseenpercn =np.true_divide(doorseenn.values, doorseenn.sum())*100
doorseenpercn
Out[120]:
In [121]:
housenumn = notsafe_subs['House number is clearly displayed'].value_counts().sort_index()
housenumpercn =np.true_divide(housenumn.values, housenumn.sum())*100
housenumpercn
Out[121]:
In [122]:
housenumvisn =notsafe_subs['House number is visible at night'].value_counts().sort_index()
housenumvisn = housenumvisn.iloc[[0,2],]
In [123]:
#housenumvisn = notsafe_subs['House number is visible at night'].value_counts().sort_index()
housenumvispercn =np.true_divide(housenumvisn.values, housenumvisn.sum())*100
housenumvispercn
Out[123]:
In [124]:
gardn = notsafe_subs['Garden kept short and neat'].value_counts().sort_index()
gardpercn =np.true_divide(gardn.values, gardn.sum())*100
gardpercn
Out[124]:
In [125]:
cameran= notsafe_subs['Security camera faces entry'].value_counts().sort_index()
campercn = np.true_divide(cameran.values, cameran.sum())*100
campercn
Out[125]:
In [126]:
dogn= notsafe_subs['I have a dog'].value_counts().sort_index()
dogpercn = np.true_divide(dogn.values, dogn.sum())*100
dogpercn
Out[126]:
In [127]:
sensorlightn= notsafe_subs['I have sensor lighting or timed lighting in my house for when I’m not at home'].value_counts().sort_index()
sensorlightpercn = np.true_divide(sensorlightn.values, sensorlightn.sum())*100
sensorlightpercn
Out[127]:
In [128]:
valuen= notsafe_subs['Valuables kept in a secured, locked safe located out of sight.'].value_counts().sort_index()
valuepercn = np.true_divide(valuen.values, valuen.sum())*100
valuepercn
Out[128]:
In [129]:
col_names = ["No",'Yes']
results = { 'Front door seen from street': doorseenpercn,
'House number is clearly displayed': housenumpercn,
'House number is visible at night': housenumvispercn,
"Doors and windows locked": doorslockedpercn,
'Electronics out of sight': computerspercn,
'Garden kept short and neat': gardpercn,
'Sensor/timed lighting when not at home': sensorlightpercn,
'Security camera at home': campercn,
'Dog at home': dogpercn,
'Valuables kept in a secured & locked safe': valuepercn}
def survey(results, col_names):
labels = list(results.keys())
data = np.array(list(results.values()))
data_cum = data.cumsum(axis=1)
category_colors = plt.get_cmap('RdYlGn')(np.linspace(0.1, 0.90, data.shape[1]))
fig, ax = plt.subplots(figsize=(9.9, 3.7))
ax.invert_yaxis()
ax.xaxis.set_visible(False)
ax.set_xlim(0, np.sum(data, axis=1).max())
for i, (colname, color) in enumerate(zip(col_names, category_colors)):
widths = data[:, i]
starts = data_cum[:, i] - widths
ax.barh(labels, widths, left=starts, height=0.9, label=colname, color=color)
xcenters = starts + widths / 2
r, g, b, _ = color
text_color = 'black'
for y, (x, c) in enumerate(zip(xcenters, widths)):
ax.text(x, y, str(int(c)), ha='center', va='center',
color=text_color)
ax.legend(ncol=len(col_names), bbox_to_anchor=(0.42, -0.15),
loc='lower left', fontsize='small')
return fig, ax
survey(results, col_names)
plt.title("House security measures (in %)")
fig.tight_layout()
plt.show()
fig.savefig('plot5.pdf')
